D6 Statistical and Prospective Analysis of Accident Risk

Topic

Beyond the monitoring of normal operations described by the two previous models, the nuclear industry and academia also employ statistical analysis methods to prospectively evaluate the likelihood and expected consequences of rare but potentially severe events—such as the residual uncertainty described in the anomaly-related model in this catalog.

This type of analysis—Probabilistic Safety Assessment—combines detailed knowledge of the potential failure modes for each component and system in a power plant (derived from accumulated industry-wide operational experience and specific engineering studies) with statistical techniques. These techniques estimate the combined probability of event sequences that might appear insignificant individually but could, in certain combinations, lead to a severe accident condition.

The interdisciplinary MIT study previously mentioned in the context of residual uncertainty serves as an example of this type of industry-wide analysis: based on projected growth of the global reactor fleet over several decades and applying the historical rate of severe accidents observed up to that point, it estimated the number of additional accidents that could reasonably be expected during that period.

Thus, this module—which begins by moving beyond routine operational monitoring—concludes by demonstrating that this prospective statistical analysis serves a distinct yet complementary function: while monitoring (stages D1 through D5) verifies in real time that an individual facility’s operations remain within normal parameters, this analysis informs regulatory and design decisions across the entire industry, guiding safety improvement efforts toward areas expected to have the greatest impact on the aggregate risk of the global fleet.

ID:3978